1 / 26

Understanding Higgs Cross Sections at LHC

Understanding Higgs Cross Sections at LHC. an overview of ATLAS effort. Jianming Qian University of Michigan (ATLAS Collaboration). Higgs Physics at the Tevatron and LHC workshop November 19, 2009, Fermilab. Higgs in ATLAS. A signal-like candidate has been seen

zavad
Download Presentation

Understanding Higgs Cross Sections at LHC

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Understanding Higgs Cross Sections at LHC an overview of ATLAS effort Jianming Qian University of Michigan (ATLAS Collaboration) Higgs Physics at the Tevatron and LHC workshop November 19, 2009, Fermilab

  2. Higgs in ATLAS A signal-like candidate has been seen in ATLAS over huge backgrounds…

  3. This talk* is about • Compiling Higgs production cross sections at LHC • estimate discovery significances; • study exclusion potential; • With real collision data • If signal, is it due to Higgs? • If not, make nice plots like this… … is not about ATLAS’s discovery potential *views expressed in this presentation are not necessarily shared by ATLAS, but they should be.

  4. ATLAS Effort • Significant theoretical and phenomenological progress in • the calculations of higher order QCD and EW corrections; • the constraining of parton distribution functions; • not an easy task to follow all the latest development for an average • analyzer… • ATLAS Higgs group conveners formed a small group to prepare Higgs • cross sections and decay branching ratios in both SM and MSSM for • first data analysis • survey theoretical calculations/tools and compare their results; • recommend a set of Higgs cross sections for the first data analysis; • build-up machineries for arbitrary LHC operation energies; • Working group members: • N. Andari, A.C. Bourgaux, M. Campanelli, G. Carrillo, M. Escalier, M. Flechl, J. Huston, • B. Mellado, J. Qian, D. Rebuzzi, M. Schram, R. Tanaka, T. Vickey, M. Warsinsky, H. Zhang, … This presentation is largely based on discussions within the working group and with theoretical colleagues that we are working with

  5. Questions from Joey • 1.a Can we have a dynamic collection of cross sections (detailing methods and • parameters used to calculate these cross sections) for all Higgs production • processes. • 1.b Understanding consistency and best use of predictions at LO, NLO, NNLO, • NLO+NLL, NNLO+NNLL • - How consistent are the predictions from CTEQ, MSTW, NNPDF ? • - How to properly include the cross section/PDF uncertainty due to • uncertainty on as? • - What is the best way of adding PDF and scale uncertainties? • - Should the factorization and renormalization scales be varied separately • or together ? • - How to deal with higher order information for differential distributions ? • - Is there a consensus on how to deal with calculations of MSSM Higgs and • their uncertainties in 4- and 5-flavor schemes ? There may not be right answers to some of these questions, but consistent/consensus approaches will be helpful for comparisons and eventual combination of results from different experiments

  6. We are examining… R. Tanaka, LAL-Orsay and D. Rebuzzi, Pavia University and INFN

  7. Gluon-Gluon-Fusion • Dominant process at LHC • exact calculations with full mt-dependence, available up to NLO; • in the large mt limit, available up to NNLO, scaling with • partial N3LO calculation large mt approximation

  8. Inclusive Calculations Converging QCD perturbative expansion: s(NkLO)  s(LO) x [ 1.0 + 0.8 + 0.3 + 0.1 + 0.05 + … ] NLO N2LO N3LO NnLL Relative corrections are weakly dependent on Higgs mass • Additional corrections/issues: • Electroweak corrections (~±5%), • bottom quark loop contribution, a reduction of ~8% at low mass; • a ~2% difference between pole and MSbar mass schemes • top mass effect, most NLO+ calculations • correct mt effect using FLO(mt): good at • low mass, but ~10% overestimate • at high mass • mt effect is small at NNLO at low mass • (Harlander et al, arXiv:0909.3420) de Florian/HIGLU - 1 Exact mt effect is known at NLO now, can we apply NLO mt correction to the large mt approximation? NLO 10 TeV

  9. Scale Uncertainties The cross section suffers large scale dependences, for example, setting mF=mR and varying between [0.5, 2]MH • However, several different approaches • in literature • different choice of the central scale; • how mF and mR are varied • Can we reach a consensus ? • bring people on the same page; • facilitate comparisons/combinations

  10. PDF and as Uncertainties • PDF: Healthy debate on how to estimate its uncertainties • general agreement on D2=1 is not a sensible option; • but no agreement on the actual 2 value • (2=100 for CTEQ and 2=50 for MSTW) • Two different philosophies on as: • CTEQ uses world average as(Mz)=0.118 at NLO in their fit; • MSTW determines as in their fit: as(Mz)=0.120 at NLO; • A ~3.5% difference if s ~ as2. de Florian et al. Outside the scale-dependence, PDF and as are the source of the next largest uncertainty If the PDF folks cannot agree, can we as experimentalists have a common approach on dealing with this ? 7 TeV

  11. Vector Boson Fusion • The process with the 2nd largest cross section • at LHC, unique event topology • With Htt, important discovery channel for • a low mass Higgs (along with Hgg) • Also s- and u-channel diagrams, • typical tag-jet requirements reduce • their contributions • Codes/calculations • VV2H (Spira): • NLO QCD, inclusive, t-channel only, no WW/ZZ interference • VBFNLO: Arnold et al., Comput. Phys. Commun. 180, 1661 (2009) • NLO QCD, exclusive • “CDD”: Ciccolini, Denner & Dittmaier, arXiv:0710.4749 • NLO QCD and EW corrections; • including t-, s- and u-channel and their interferences

  12. Vector Boson Fusion • On the same basis, results from different calculations agree within • 1% (Les Houches 2007 report, arXiv:0803.1154); • NLO corrections are relatively small (~+5% for QCD, ~-5% for EW); • Relative small scale and PDF uncertainties (anti)diagonal between [1/4, 4]MW All these calculations are done with narrow width approximation, their validity at high mass is not assured…

  13. WH and ZH • gold at Tevatron, more like brass at LHC though • boosted Higgs might improve sensitivity • s ~ 1 pb for mH=120 GeV at s=10 TeV • one of the few channels for potential Yukawa • coupling measurements for a low mass Higgs • Calculations up to NNLO in QCD and NLO in EW • NLO calculations • V2HV (Spira): public • MCFM (Campbell et al): public • NNLO calculations: private • Brein, Djouadi & Harlander, PLB 579, 149 (2004) • Ciccolini, Dittmaier & Kramer, PRD68, 073003 (2003) • Brein et al., hep-ph/0402003 Cross section numbers are provided by Harlander with the latest PDF (MSTW2008)

  14. WH and ZH • NLO QCD corrections increase the cross section by +30%, • - new qgq(W/Z)H process • NLO EW corrections reduce it by up to 7% • NNLO QCD corrections • - small increase (~2%) for WH • - significant increase (~10%) for ZH because of ggZH contribution WH scale variation • scale variation reduces from • ~10% at NLO to <3% at NNLO • At NLO, PDF uncertainty ~5%. • The uncertainty at NNLO is • being evaluated. The inclusive cross section is precise enough for any practical purpose Brein, Djouadi and Harlander: hep-ph/0307206: scale varied one at a time in between [1/3,3]MHV

  15. ttH Production • possibility to measure Htt Yukawa coupling • s ~ 0.4 pb for mH=120 GeV at s=10 TeV, • large background, small S/B ratio (Dawson et al.) • two independent QCD NLO calculations • Dawson et al. PRD68, 034022 (2003) • Beenakker et al., Nucl. Phys. B653, 151 • NLO K-factor: 1.1-1.4 depending on scale • Codes: • Public: HQQ (Spira), but only in LO • private NLO available • Remaining uncertainty • ~15% from scale variation • ~5% from PDF Updated NLO numbers from Michael Spira, the real issue is background such as tt+jets…

  16. MSSM Higgs Sector • Five Higgs bosons: h0, H0, A0, H+, H- • Two neutral Higgs bosons are usually • degenerate in mass  f • Can be parameterized by two parameters • (tanb, mA) in benchmark scenario, assumed • mh-max scenario in our study • Enhanced coupling of neutral Higgs to b quark •  new b-associated production ggfbb, dominant for tanb>10 •  increased b-loop contribution for ggf, dominant for tanb<10 • Both HDECAY/HIGLU (Spira) and FeynHiggs (Heinemeyer et al) calculate • mass spectrum, decay branching ratios and production cross sections.

  17. MSSM gg • Similar to SM ggH production • modified couplings to top and b quark loops; • enhanced b-quark loop contribution at large tanb • Proposal by Spira to reweight SM cross sections with coupling ratios • to take advantage of more precise calculations of SM processes • (only works for scalar, NNLO calculation • unavailable for pseduoscalar) • Ongoing work • reweighting implementation, comparison with FeynHiggs calculations; • understand scale and PDF uncertainties; • squark-loop in HIGLU being implemented (Spira) t-quark t-b b-quark loop interference loop

  18. MSSM ggbb • Two schemes for cross section calculations: • 4-flavor scheme: ggbbH, up to NLO in QCD • b-quark from gluon splitting, large log contribution • Dawson et al., PRD69, 074027 (2004) • Dittmaier, Kramer & Spira, PRD70, 074010 (2004) • Private code by Kramer and Spira • 5-flavor scheme: bbH, up to NNLO in QCD, b-quark from PDF • Harlander & Kilgore, PRD 68, 013001 (2003) • Public code: bbh@nnlo MRST2004 Two calculations differ by ~15%, but agree within uncertainties

  19. MSSM Charged Higgs • Two main production processes • top pair production with tHb, dominant for mH<<mt; • associated production , dominant for heavy Higgs • ggtbH- (4-flavor) or gbtH- (5-flavor) • Approximate NNLO s(tt) calculation • Lagernfeld, Moch & Uwer, arXiv:0907.2527 • Associated production • 5-flavor calculation and semi-private code • from T. Plehn • - 4-flavor calculation • from Dittmaier et al. arXiv:0906.2648 • Br(tbH) and Br(H…) from FeynHiggs Dittmaier et al • Uncertainties: • Scale: ~6% (20%) for light (heavy) Higgs • PDF: ~5% A factor of 1.4 difference!

  20. Event Kinematics • High order corrections can also impact event kinematics. For ggH • give Higgs a pT kick and therefore alter pT/h distributions and angular • correlations of particles from Higgs decays • increase jet activity in the event • Higgs decay kinematics depends on its pT and rapidity. Both • FEHiP and HNNLO programs calculate differential cross sections • significant change in the low pT spectrum (further modified by NNLL) • negligible change in the rapidity distribution

  21. Selection Efficiencies • Event kinematics change the most from • LO to NLO, small change from NLO to • NNLO, no significant changes are expected • from even higher order corrections • With mild kinematic cuts, higher order • corrections have relatively small effects • on selection efficiencies

  22. HWW*ll: Jet Veto However… • Event jet activity: • difficult to model it well in MC; • strongly impacted by high order • corrections; • significant decrease in jet veto • surviving probability at h.o. The (parton) jet veto kills all the gain from the h.o. corrections! • 1-jet bin has ~30% of the signal, • can probably be recovered; • 2-jet bin has ~10% of the signal, • they are probably lost Effects of soft radiations need to be understood

  23. Comments on Backgrounds • Our work has so far focused on signal processes. For many processes, • signal cross section is less an issue compared with that of background • process. A few examples… • ggHVV • How large is the ggVV NLO correction to the SM VV cross section? • Will the correction impact event topology ? • VH, ttH • these analyses are overwhelmed by backgrounds, signal cross sections • are not real issues. Kinematic regions with highly boosted Higgs will • significantly improve S/B, but it places a premium on the understanding • of exclusive distributions of both signal and backgrounds.

  24. Help Needed… • Make your code public and interface it to LHAPDF ! • Can experiments agree to quote only published numbers or those • from publically available codes? • one of the most difficult issues is to deal with private codes • and/or calculate with latest PDF… • Consensus on scale choices • Central scales and ranges for scale variations for individual processes • Agreement on procedures for estimating PDF (as ?) uncertainties • among PDF communities? Tevatron and LHC communities ? • Interface NkLO calculations with parton shower MC generators, • resolve the difference between 4- and 5-flavor calculations of • MSSM Higgs • Others… • Common SM parameter setup/values, e.g. mb(mb) in MSbarvsmb pole • mass; NLO top mass effect for ggH; Decay width effect for a heavy Higgs; • Interferences between Higgs and SM background diagrams; …

  25. Summary • An active ATLAS group surveying theoretical calculations/tools, • comparing results from different tools and compiling the best • estimates of Higgs cross sections; • Have been actively collaborating with our theoretical colleagues • to understand differences and limitations of different calculations; • Consensuses on a number of issues among theory and PDF • communities will be very helpful; • Nothing of these is ATLAS specific. A common approach between • ATLAS and CMS will be beneficial for potential future combination • of results.

  26. ggF PDF

More Related